Milestones
List view
<!-- eech-native-roadmap:v1:M8 --> ## Intent Compare matched scenarios and calibrate legitimate host-dependent differences. Seek comparable operational behaviour, not identical trajectories, exact casualty totals or a predetermined winner. ## Acceptance gate - Agree comparable scenario families, operational measures and variation criteria before tuning. - Classify differences as integration defect, campaign discrepancy, physical difference or explicit scenario/tuning choice. - Calibrate only justified causes while keeping campaign policy shared and preserving each host baseline. - Use repeated comparisons, tests, structured results and Tacview to accept a versioned cross-host baseline. ## Exit question Can we justify the remaining differences and demonstrate the intended campaign dynamics without disguising defects? ## Baseline protected A reviewed cross-host baseline, tuning provenance, comparison criteria and accepted residual differences. **Not sufficient:** Matching one total, forcing a familiar winner, weakening the reference scenario or balancing away missing observations. Acceptance follows [M7](https://github.com/flying-dice/eech_source_code/issues/43); implementation investigations may overlap. No dates or implementation changes are prescribed. Close only after an explicit evidence review. Epic: https://github.com/flying-dice/eech_source_code/issues/47 Project: https://github.com/orgs/flying-dice/projects/8
No due date•0/4 issues closed<!-- eech-native-roadmap:v1:M7 --> ## Intent Demonstrate the intended deployment path in the real DCS runtime. DCS supplies the physical world; the DLL supplies campaign decisions and progression. ## Acceptance gate - Demonstrate loading and supported lifecycle in the actual DCS-owned Lua environment. - Show real campaign requests causing DCS activity and real outcomes affecting subsequent campaign decisions. - Exercise representative adverse paths and retain correlated Tacview and structured evidence. - Characterise runtime stability/performance and accept a sustained operating baseline while preserving EECH World. ## Exit question Is the campaign genuinely operating in DCS, rather than only building or loading successfully? ## Baseline protected A functioning, versioned DCS operating baseline with measured behaviour and known differences. **Not sufficient:** A Windows build, successful require, initial spawn or test-script campaign orchestration. Acceptance follows [M6](https://github.com/flying-dice/eech_source_code/issues/39); implementation investigations may overlap. No dates or implementation changes are prescribed. Close only after an explicit evidence review. Epic: https://github.com/flying-dice/eech_source_code/issues/43 Project: https://github.com/orgs/flying-dice/projects/8
No due date•0/4 issues closed<!-- eech-native-roadmap:v1:M6 --> ## Intent Build our integration with DCS, not DCS itself. Translate supported requests and observations while making non-equivalent capabilities and approximations explicit. ## Acceptance gate - Establish evidenced DCS mappings and the supported capability scope. - Provide working adapter interactions and lifecycle with observable failures. - Verify contract semantics and distinguish controlled/captured evidence from actual DCS execution. - Retain the corresponding EECH World baseline and document readiness/prerequisites for live operation. ## Exit question Is the DCS adapter coherent and ready to exercise the same campaign product in the actual host? ## Baseline protected Accepted adapter behaviour and shared contract, alongside the reference-host baseline. **Not sufficient:** Plausible API mappings, another standalone simulator or mocked tests presented as live-DCS proof. Acceptance follows [M5](https://github.com/flying-dice/eech_source_code/issues/35); implementation investigations may overlap. No dates or implementation changes are prescribed. Close only after an explicit evidence review. Epic: https://github.com/flying-dice/eech_source_code/issues/39 Project: https://github.com/orgs/flying-dice/projects/8
No due date•0/4 issues closed<!-- eech-native-roadmap:v1:M5 --> ## Intent Resolve the difference between a whole-engine Lua wrapper and a reusable campaign module. Establish real world authority at the host boundary while preserving the accepted EECH World behaviour. ## Acceptance gate - Evidence the shared semantic contract: ownership, identity, time, observations, action outcomes and lifecycle. - Demonstrate that host-supplied physical outcomes drive the campaign without hidden shared-world authority. - Preserve the M4 baseline across boundary changes, investigating and approving every meaningful difference. - Verify material DCS compatibility assumptions with appropriate probes; keep remaining uncertainty explicit. ## Exit question Can the physical world be replaced without replacing campaign policy, and has the reference baseline survived? ## Baseline protected A host-separated reference baseline and an evidenced common host contract. **Not sufficient:** Renamed crates, interface declarations or unchanged recordings over hidden co-resident world state. Acceptance follows [M4](https://github.com/flying-dice/eech_source_code/issues/31); implementation investigations may overlap. No dates or implementation changes are prescribed. Close only after an explicit evidence review. Epic: https://github.com/flying-dice/eech_source_code/issues/35 Project: https://github.com/orgs/flying-dice/projects/8
No due date•8/9 issues closed<!-- eech-native-roadmap:v1:M4 --> ## Intent Strengthen existing tests and observations into credible regression evidence. Testing begins in M1; M4 establishes the confidence required for larger refactoring. ## Acceptance gate - Map supported capabilities and risks to public-boundary behavioural checks, including important adverse paths. - Demonstrate meaningful regression sensitivity and characterise expected variation. - Retain lower-level conformance without overstating its scope or independence. - Establish repeatable acceptance, retained artifacts and explicit review of baseline changes. ## Exit question Could we change the implementation substantially and credibly detect damaged campaign behaviour or host interactions? ## Baseline protected An accepted regression portfolio with source/expectation provenance and visible blind spots. **Not sufficient:** A coverage percentage, crash-free soak, final object count or regenerated expected output alone. Acceptance follows [M3](https://github.com/flying-dice/eech_source_code/issues/27); implementation investigations may overlap. No dates or implementation changes are prescribed. Close only after an explicit evidence review. Epic: https://github.com/flying-dice/eech_source_code/issues/31 Project: https://github.com/orgs/flying-dice/projects/8
No due date•8/8 issues closed